648

Transition Metal Chemistry 24: 648±654, 1999.

Ó 1999 Kluwer Academic Publishers. Printed in the Netherlands.

The existence of ferrous ferricyanide Edilso Reguera* Esc. Sup. de FõÂsica y Matem.-IPN, Unidad Prof. ``ALM'', Colonia Lindavista, 07738 MeÂxico DF, MeÂxico Jose FernaÂndez-BertraÂn Centro de QuõÂmica FarmaceuÂtica, La Habana, Cuba Jorge Balmaseda Centro Nacional de Investigaciones Cientõ®cas, La Habana, Cuba Received 25 January 1999; accepted 22 February 1999

Abstract Evidence in the literature on the synthesis of ferrous ferricyanide is critically discussed. Pyrolysis and pressure e€ects on Prussian Blue lead to ferrous ferrocyanide together with decomposition by-products, and not to ferrous ferricyanide. The latter compound could be a precursor in the formation of Turnbull's Blue or an excited state of Prussian Blue, but it is not a stable chemical species. Introduction The iron hexacyanoferrates have been an exhaustively studied family of complexes. The best known and most stable member is ferric ferrocyanide or Prussian Blue (PB). Ferric ferricyanide, Prussian Brown (PBr), is an unstable product which, in air, passes through a green phase, Prussian Green (PG), and ends up in PB. PG is not a de®nite compound but a mixture of PBr and PB [1, 2]. Ferrous ferrocyanide, Williamson's White (WW), is a white product that oxidizes rapidly in air to PB [3, 4]. When solutions of Fe2+ and [Fe(CN)6]3) are mixed, the insoluble Turnbull's Blue (TB) precipitates. TB is not ferrous ferricyanide but a ferric ferrocyanide, similar but not identical to PB [1]. The elusive ferrous ferricyanide cannot be detected in the reduction of PBr [1] or in the oxidation of WW or PB [1, 2]. However, many researchers have endeavored to obtain this (notorious) species [5±10]. In this paper we present a critical review of the published results, together with new experimental evidence on this subject, and arrive at the conclusion that ferrous ferricyanide does not exist as a stable chemical species. Experimental Samples of PB, PBr, WW and TB were prepared as described elsewhere [1, 2, 4] and then aged in order to obtain di€erent species of ferric ferrocyanide to be pyrolyzed in vacuum. To minimize the oxidation of WW during its preparation, the mechanochemical reaction [4] was carried out in a glove box containing Ar (99.99%). * Author for correspondence

The studied compositions correspond to insoluble Prussian Blue (Fe4[Fe(CN)6]3, á xH2O), soluble Prussian Blue (FeK[Fe(CN)6] á xH2O), aged Prussian Brown, (Fe (H3O)[Fe(CN)6] á xH2O) and Turnbull's Blue (Fe3(OH) [Fe(CN)6]2 á xH2O), labeled, IPB, SPB, APBr and TB, respectively. These complexes correspond to the samples Nr. 1, 2, 11 and 23 in a previous work [1], whose nature and purity were established by chemical analysis, i.r. and MoÈssbauer spectroscopies, and X-ray powder di€raction (XRD). The expected MoÈssbauer spectrum for a stoichiometric ferrous ferricyanide (``Fe3[Fe(CN)6]2 á xH2O'') was experimentally simulated from a mixture of nickel ferricyanide (Ni3[Fe(CN)6]2 á xH2O) and ferrous cobalticyanide (Fe3[Co(CN)6]2 á xH2O) in a 2:3 molar ratio. This mixture and in this molar ratio guarantee the expected low spin iron(III) and high spin iron(II) populations corresponding to a hypothetical stoichiometric ferrous ferricyanide, and in the appropriate structural environments. Nickel ferricyanide and ferrous cobalticyanide were prepared and characterized as previously described [11, 12]. The vacuum pyrolysis of IPB, SPB, APBr and TB samples was carried out at 10)6 Torr. during 1 h heating at 250, 350 and 400 °C in sealed glass tubes. To prevent their oxidation the pyrolyzed samples were handled in an Ar atmosphere using a glove box and were then preserved in Nujol. The pyrolysis products obtained in this manner were characterized by i.r. and MoÈssbauer spectroscopies. The i.r. spectra were recorded in Nujol using an FT-IR spectrometer (Atti Matson, Genesis Series). The MoÈssbauer spectra were recorded at room temperature with a 57Co in Rh in source, using a constant acceleration spectrometer. All spectra were ®tted using an iterative least squares minimization

649 algorithm and pseudo-Lorentzian line shapes to obtain the values of isomer shift (d), quadrupole splitting (D), linewidth (G) and relative area (A). The isomer shift values are reported relative to metallic iron (a-Fe) at room temperature. Results and discussion Expected MoÈssbauer and i.r. spectra for ferrous ferricyanide The postulated chemical species, ferrous ferricyanide, belongs to the family of divalent transition metal hexacyanometallates(III). These compounds normally have a cubic structure with ao close to 10 AÊ [13±15]. In accord with their stoichiometry, M03 ‰MIII …CN†6 Š2
xH2 O, one third of the [MIII(CN)6] anion groups are left vacant, creating a rather open structure and also a certain local asymmetry [14, 15]. Related to that local asymmetry is the well resolved doublet observed for lowspin iron(III) cations in the MoÈssbauer spectra of divalent transition metal ferricyanides [11] (see Figure 1a and Table 1). The vacant [MIII(CN)6] positions are ®lled with water molecules, which act as ligands completing the coordination sphere of half of the outer cations (M0 2‡ ) and as zeolitic water. As a consequence, the outer cation in these compounds is found in two di€erent structural sites, one with a coordination sphere of six CN ligands at N ends and one with a mixed coordination sphere of CN and aquoligands. This is observed in the MoÈssbauer spectra of ferrous hexacyanometallates(III) as two well resolved high-spin iron(II) doublets [16] (see Figure 1b and Table 1). According to this structural model, the expected MoÈssbauer spectra for a hypothetical ferrous ferricyanide, ``Fe[Fe(CN)6]2 á xH2O'', must be composed of three quadrupole doublets, two for Fe2+ cations and one for FeIII cations (see Table 1). Figure 1c shows the expected MoÈssbauer spectrum for this compound. As already pointed out, that spectrum was obtained from a 2:3 molar ratio of nickel ferricyanide and ferrous cobalticyanide. An important and additional characteristic of the expected MoÈssbauer spectra of ferrous ferricyanide is its temperature dependence. The quadrupole splitting (D) of iron in low-spin iron(III) and high-spin iron(II)

Fig. 1. MoÈssbauer spectra at room temperature of: a) Nickel ferricyanide (Ni3[Fe(CN)6]2 á xH2O); b) Ferrous cobalticyanide (Fe3[Co (CN)6]2 á xH2O); c) Mixture of Ni3[Fe(CN)6]2 á xH2O and Fe3[Co (CN)6]2 á xH2O in a 2:3 molar ratio. The MoÈssbauer spectrum of this mixture must be similar to that expected for an hypothetical ferrous ferricyanide, ``Fe3[Fe(CN)6]2 á xH2O''.

electronic con®gurations is mainly due to the relative electronic population in the t2g levels which changes with the temperature of the MoÈssbauer measurement [17]. On lowering this temperature, the D value for both low spin iron(III) and high spin iron(II), must increase since the higher t2g sub-levels depopulate and the electronic population around the cations becomes more asymmetric. The relative electronic population of t2g

Table 1. MoÈssbauer parameters at room temperature of nickel ferricyanide, ferrous cobalticyanide and a 2:3 molar ratio of nickel ferricyanide and ferrous cobalticyanide* Sample

d** (mm s)1)

D (mm s)1)

G (mm s)1)

A (%)

Assignment

Nickel ferricyanide (Ni2+ FeIII) Ferrous cobalticyanide (Fe2+ CoIII)

)0.15 1.13 1.14 1.13 1.14 )0.15

0.49 1.90 1.09 1.90 1.08 0.48

0.34 0.45 0.50 0.46 0.51 0.36

100 52 48 32 30 38

Low-spin iron(III) High-spin iron(II) High-spin iron(II) High-spin iron (II) High-spin iron (II) Low-spin iron (III)

Mixture of Ni2+FeIII and Fe2+CoIII in a 2:3 molar ratio

* Errors in isomer shift (d), quadrupole splitting (D), and linewidth (G) are no higher than 0.01 mm s)1. ** The values of d are reported relative to a-Fe.

650 sub-levels can also be changed, and in consequence the value of D, by removing or adding water molecules or gegenions located close to the t2g orbitals [11, 18]. Dehydration of ferricyanides, for instance, produces a signi®cant increase in the measured value of D [11]. I.r. spectra provide a simple and fast method for identifying ferrocyanides and ferricyanides. The CN stretching vibration m(CN), is an excellent sensor of the oxidation state of the inner cation [iron(II) and iron(III)]. Ferrocyanides absorb in the 2000±2100 cm)1 region [2, 19, 20], while ferricyanides exhibit this band at ca. 80 cm)1 higher [2, 12, 21]. An increase in the valence of the outer cation has a smaller but detectable e€ect, increasing m(CN) by 10±15 cm)1 [4, 21]. Ferric ferricyanide (PBr), for instance, absorbs at 2172 cm)1 [2], ferric ferrocyanide (PB) at 2080 cm)1 [2, 20] and ferrous ferrocyanide (WW) at 2068 cm)1 [4, 22]. As regard intensities, the ferrocyanide bands are very intense and broad, whereas ferricyanide bands are narrow and of medium intensity. From these facts, the m(CN) value for ferrous ferricyanide can be predicted to be a narrow medium intensity band at ca. 2160 cm)1 (see Figure 2). This prediction can also be made from correlations of m (CN) frequencies and from the analogies of ferrocyanides and cobalticyanides as well as mCN di€erences between Fe2+ and Fe3+ hexacyanometallates [12, 19, 23].

High pressure e€ect on Prussian Blue Drickamer et al. [5, 24] have made extensive studies of hexacyanoferrates under pressure. They have reported the MoÈssbauer spectra of PB at di€erent pressures and temperatures, using 57Fe selectively labeled outer and inner cations [5]. The MoÈssbauer spectra reported give conclusive evidence that the outer high-spin Fe3+ cation is reduced to high spin Fe2+ cation at high pressure and temperature, up to 90% conversion. The spectra of the inner low spin 57FeII is not as conclusive. They report the formation of ferrous ferricyanide under pressure supposedly through an electron transfer from the inner cations to the outer ones. However, we interpreted the same spectra as a main absorption due to low-spin iron(II) and a less abundant absorption due to high-spin Fe2+ cations. This interpretation is in accord with the reported behavior of other ferrocyanides at high pressures [5] where the MoÈssbauer absorption from low-spin iron(II) appears as a quadrupole splitting doublet and not as the usual single line at room temperature and atmospheric pressure [11, 20]. Walley et al. [25, 26] have shown that the reduction of ferricyanides to ferrocyanides under pressure is not due to a reversible pressure e€ect but to a shearing e€ect which causes mechanochemical reduction and partial destruction of the hexacyanoferrate anion with production of cyanide radicals, which can have a reducing e€ect on the system. Mechanochemical reduction of ferricyanides to ferrocyanides on milling takes place even at atmospheric pressure [27]. From these reasons, we believe that the observed e€ect of pressure on PB is the formation of ferrous ferrocyanide, WW, without the presence of ferrous ferricyanide. Vacuum pyrolysis of Prussian Blue

Fig. 2. Characteristic CN stretching absorption bands of iron hexacyanoferrates: Fe2+FeII (ferrous ferrocyanide, WW, with a shoulder due to partial oxidation and formation of ferric ferrocyanide); Fe3+FeII (ferric ferrocyanide, PB); Fe3+FeII (ferric ferricyanide, PBr); Fe2+FeIII (expected CN absorption band for the postulated ferrous ferricyanide).

Collins et al. [6±10] have prepared ferrous ferricyanide by vacuum pyrolysis of PB. They have measured the MoÈssbauer and i.r. spectra of the product and ascribed the stability of ``ferrous ferricyanide'' to the loss of water molecules from the crystal structure, which causes its contraction. This would be equivalent to a pressure e€ect as studied by Drickamer et al. [5, 20]. The i.r. spectra of ferrous ferricyanide is reported as ``similar but not identical'' to those of PB [9], which has a broad intense band at 2080 cm)1 [2, 20]. However, the m(CN) band of ferrous ferricyanide should be quite di€erent, as already discussed above (see Figure 2). The only i.r. spectrum similar to PB is that of WW with an intense broad absorption at 2068 cm)1 [4]. The reported behavior of the pyrolyzed product, which returns in days to PB, when exposed to the atmosphere, is in accordance with its nature as WW [4], not as ferrous ferricyanide. We have pyrolyzed PB samples and always the resulting i.r. spectra shown a main absorption at 2068 cm)1 indicating the formation of ferrous ferrocyanide and a shoulder ca. 2080 cm)1. This last shoulder must be due to the non transformed fraction of PB during the

651 pyrolysis process or due to a partial oxidation during the handling of the pyrolyzed samples (see Figure 3). No i.r. absorption bands were observed in the 2160 cm)1 region corresponding to ferricyanides. From the reported MoÈssbauer spectra of pyrolyzed PB samples [6±10] it is evident the reduction of the outer Fe3+ to Fe2+. MoÈssbauer spectra of the labeled 57Fe3+ PB samples, which initially is a high-spin ferric doublet, after pyrolysis at 400 °C shows intense absorptions due to high-spin Fe2+ [8]. Conclusive evidence for the nature of the low spin iron(II) or iron(III), in the pyrolysis product, can be reached from the excellent series of published MoÈssbauer spectra at 77 and at 297 K [9, 10]. No signi®cant changes are observed in the pro®le of MoÈssbauer absorption corresponding to the inner cation at these two temperatures. Of these two electronic con®gurations, only the low spin iron(II) one does not signi®cantly change its MoÈssbauer parameters with the measurement temperature. In consequence, we have ignored the presence of signals from low-spin iron(III) in these MoÈssbauer spectra. The possibility of an unresolved low-spin iron(III) quadrupole doublet for the pyrolyzed PB samples is also discarded since, in addition to the e€ect of the temperature of measurement, the pyrolysis process introduces considerable dehydration of the sample which signi®cantly increases the mea-

sured D value for divalent transition metal ferricyanides [11]. MoÈssbauer spectra of our pyrolyzed PB samples can be ®tted as a superposition of a singlet of low-spin iron(II), two doublets of high spin iron(II) and, in some cases, a small contribution of a high-spin iron(III) doublet from the fraction of non transformed sample (see Table 2). For SPB, one of the high-spin iron(II) doublets has MoÈssbauer parameters similar to those observed in a pure potassium-ferrous ferrocyanide, WW (see Figure 4 and Table 2). The other high-spin iron(II) doublet could be due to a decomposition by-product and perhaps to a certain contribution of another species of ferrous ferrocyanide with MoÈssbauer parameters di€erent to those of potassium-ferrous ferrocyanide. This last possibility is supported by the observed area ratio of high-spin iron(II)/low-spin iron(III) (see Table 2). According to the chemical composition of the SPB sample studied, FeK[Fe(CN)6] á xH2O, it is impossible to obtain a pure WW, even in a reduction process, without occurrence of partial decomposition of the sample during the heating, introducing, for instance, an external reducing agent, like H2. A similar analysis can be made for the other PB species (IPB, APBr and TB) considered in this study, but the MoÈssbauer parameters of the resulting pyrolyzed products are

Fig. 3. I.r. spectra (CN stretching region) of: (a) Ferrous ferrocyanide (WW), (3); (b) Ferric ferricyanide (PBr), (1), partially reduced to ferric ferrocyanide (APBr), (2); (c) Pyrolyzed SPB; (d) Pyrolyzed APBr; (e) Pyrolyzed IPB; (f) Pyrolyzed TB. These IR spectra correspond to PB samples pyrolyzed in vacuum at 400 °C during one hour. In addition to the main absorption at 2068 cm)1 (3) due to ferrous ferrocyanide, the pyrolyzed samples also shown a shoulder at about 2080 cm)1 (2) indicating presence of ferric ferrocyanide.

652 Table 2. MoÈssbauer parameters at room temperature of ferrous and ferric ferrocyanide samples and their products of pyrolysis in vacuum at 400 °C during one hour* Sample

d** (mm s)1)

D (mm s)1)

G (mm s)1)

A (%)

Assignment

WW

1.07 )0.09 0.41 )0.13 1.02 1.16 0.41 )0.14 0.39 )0.14 1.03 1.10 0.39 )0.14 0.38 )0.15 1.01 1.08 0.38 )0.14 0.37 )0.15 1.04 1.09 0.37 )0.14

0.45 ± 0.27 ± 1.87 0.39 0.28 ± 0.33 ± 1.94 0.46 0.33 ± 0.51 ± 1.95 1.07 0.51 ± 0.53 ± 2.01 0.83 0.53 ±

0.59 0.21 0.48 0.30 0.39 0.37 0.48 0.40 0.40 0.37 0.53 0.49 0.40 0.36 0.43 0.30 0.52 0.63 0.43 0.45 0.44 0.28 0.47 0.50 0.44 0.38

49 51 50 50 31 24 11 34 51 49 63 17 5 41 57 43 28 26 8 38 60 40 32 23 9 36

High-spin iron(II) Low-spin iron (II) High-spin iron(III) Low-spin iron(II) High-spin iron(II) High-spin iron(II) High-spin iron(III) Low-spin iron(II) High-spin iron(III) Low-spin iron(II) High-spin iron(II) High-spin iron(II) High-spin iron(III) Low-spin iron(II) High-spin iron(III) Low-spin iron(II) High-spin iron(II) High-spin iron(II) High-spin iron(III) Low-spin iron(II) High-spin iron(III) Low-spin iron(II) High-spin iron(II) High-spin iron(II) High-spin iron(III) Low-spin iron(II)

SPB Pyrolyzed SPB *** APBr Pyrolyzed APBr *** IPB Pyrolyzed IPB *** TB Pyrolyzed TB ***

* Errors in isomer shift (d), quadrupole splitting (D), and linewidth (G) are no higher than 0.01 mm s)1. ** The values of d are reported relative to a-Fe. *** The values of d, D, G of the low intensity doublet of Fe3+ were constrained to their values in the starting PB samples.

slightly di€erent to those of SPB (see Table 2). As regard pyrolysis temperatures, below 300 °C the transformation rate is very low and above 450 °C a large fraction of the studied sample decomposes. The reduction of the outer ferric cation in PB on pyrolysis could be related to a partial decomposition of the initial compound on heating, releasing CN radicals, which is a well-known phenomena in PB [10, 29, 30]. The formation of ferrous ferrocyanide in these conditions can be ascribed to the reduction of Fe3+ by the CN radicals with formation of C2N2. The occurrence of decomposition during the pyrolysis of PB is evident from the following facts: (1) after pyrolysis in the sealed glass tube a large amount of C2N2 is detected; (2) the observed ratio of high spin iron species to low spin ones increases on pyrolysis (see Table 2); (3) when pyrolyzed PB samples are re-oxidized in air, always a high spin iron(III) by-product is obtained. This last ferric phase has also been observed during the vacuum pyrolysis of other hexacyanoferrates, and, tentatively identi®ed according to its MoÈssbauer parameters, as an iron oxide of small particle size [11, 27]. Inoue et al. [28] have studied the e€ect of vacuum pyrolysis of PB followed by quenching at liquid nitrogen temperature. For samples pyrolyzed above 250 °C the resulting product was reported to be amorphous according to its XRD pattern. The published MoÈssbauer spectrum for this amorphous product shows absorptions of high spin iron(II), due to reduction of Fe3+, plus a

doublet in the region of low spin iron. This last doublet does not mean presence of low spin iron(III) since amorphous or disordered ferrocyanides can give a partially resolved doublet for low spin iron(II) [27]. As already discussed, metal ferrocyanides decompose at high temperatures loosing CN ligands [27]. A random loss of these ligands must destroy the skeletal structure of the complex giving an amorphous material. From the experimental evidence discussed above, the formation of ferrous ferricyanide during the vacuum pyrolysis of ferric ferrocyanide is discarded. MoÈssbauer and i.r. spectra of pyrolyzed PB samples correspond to a mixture of ferrous ferrocyanide and other by-products, not to ferrous ferricyanide. Ferrous ferricyanide as a precursor of Turnbull's Blue Turnbull's Blue can be obtained by mixing solutions of soluble Fe2+ and [Fe(CN)6]3) salts. The same stoichiA) [FeII(CN)6]2 á ometry is always obtained, Fe3‡ 3 ) xH2O, where A is an accompanying anion to balance the charge [1]. The FeII:Fe3+ ratio of 2/3 of TB is di€erent from that of IPB, 3/4 or SPB, 1/1 [1]. This ratio gives an important clue as to the nature of the precursor species of TB. As discussed by the authors [1, 2], the initial precipitation of Fe2+ and [Fe(CN)6]3) ions leads to an unstable ferrous ferricyanide species which determines the stoichiometry of TB [1], Fe2‡ 3 [FeIII(CN)6]2. This precursor is unstable as the reducing

653 hv

FeII …CN†6 Fe3‡ ÿ! FeIII …CN†6 Fe2‡ ground state excited state

…3†

Due to the electron transfer caused by the electromagnetic interaction with the ground state, the excited singlet resembles a ferrous ferricyanide species. Since the transition is very intense, it is strongly allowed and the life time of the excited state is short, probably less than 10)9 s, returning to the ground state by ¯uorescence. That is why it is impossible to have a stable, long lived, ferrous ferricyanide chemical species. Conclusions III The chemical species Fe2‡ 3 ‰Fe …CN†6 Š2 , ferrous ferricyanide, has not been detected in the oxidation of WW or in the reduction of PBr. The species reported from PB samples under high pressures is probably WW. The vacuum pyrolysis of PB leads to WW and not to ferrous ferricyanide. The only evidence of this species is as a precursor state in the synthesis of TB or as an excited electronic state of PB.

Fig. 4. MoÈssbauer spectra at room temperature of: (a) Soluble Prussian Blue, FeK[Fe(CN)6] á xH2O; (b) FeK[Fe(CN)6] á xH2O pyrolyzed in vacuum at 400 °C during one hour; (c) Potassium-ferrous ferrocyanide, FeK2[Fe(CN)6] á xH2O.

Fe2+ cation and the oxidizing [Fe(CN)6]3) anion have an open path for electron transfer through the CN bridge [1]. This fast internal process leads to the mixed valence ferrous-ferric ferrocyanide system according to Equation (1): 3ÿ III Fe2‡ 3 ‰Fe …CN†6 Š2

ÿ! Fe

2‡

4ÿ II Fe3‡ 2 ‰Fe …CN†6 Š2

…1†

The mixed valence species can be oxidized by dissolved air or by ferricyanide in solution to a charged ferric ferrocyanide species: O2 or ‰FeIII …CN†6 Š3ÿ

4ÿ II Fe2‡ Fe3‡ ƒƒƒƒƒƒƒƒƒƒ! 2 ‰Fe …CN†6 Š2 4ÿ 1‡ II fFe3‡ 3 ‰Fe …CN†6 Š2 g

…2†

Acquisition of an anion from solution to balance the ÿ II charge leads to TB, Fe3‡ 3 A ‰Fe …CN†6 Š2 . TB com) 2) pounds with A = Cl , 1/2(SO4) and OH) have been synthesized [1]. Ferrous ferricyanide as an excited state of Prussian Blue The intense blue of PB is due to promotion of an electron from the central low-spin iron(II) cation to the outer high spin iron(III) cation. This electronic transition is called a ``charge transfer'' band [31, 32]. The electronic process can be depicted as:

Acknowledgements The authors thank L. NunÄez for help in preparing the manuscript, and C. Diaz and D. Fernandez for their initial cooperation in this work. References 1. E. Reguera, J. FernaÂndez-BertraÂn, A. Dago and C. DõÂ az, Hyperf. Interact., 73, 295 (1992). 2. E. Reguera, J. FernaÂndez-BertraÂn, C. DõÂ az and J. Molerio, Hyperf. Interact., 73, 285 (1992). 3. A.G. Sharpe, in The Chemistry of Cyano Complexes of the Transition Metals, Academic Press, London (1976) Ch. VII. 4. E. Reguera, J. FernaÂndez-BertraÂn and L. NunÄez, Z. Naturforsch., 50B, 1067 (1995). 5. S.C. Fung and H.G. Drickamer, J. Chem. Phys., 51, 4353 (1969). 6. D.S. Murty and R.L. Collins, Bull. Amer. Phys. Soc., Ser II, 15, 819 (1970). 7. R.L. Collins, D.S. Murty and J.G. Cosgrove; Bull. Amer. Phys. Soc., Ser II, 16, 23 (1970). 8. D.S. Murty, J.G. Cosgrove and R.L. Collins, Nature, 231, 311 (1971). 9. J.G. Cosgrove, R.L. Collins and D.S. Murty, J. Amer. Chem. Soc., 95, 1083 (1973). 10. R. Robinette and R.L. Collins, J. Coord. Chem., 4, 65 (1974). 11. E. Reguera and J. FernaÂndez-BertraÂn, Hyperf. Interact., 88, 49 (1994). 12. J. FernaÂndez-BertraÂn, J. Blanco and E. Reguera, Spectrochim. Acta, 46A, 685 (1990). 13. A. Ludi and H.U. Gudel, Structure and Bonding, 14, 1 (1973). 14. G.W. Beall, W.O. Milligan, J. Korp and I. Bernal, Inorg. Chem., 16, 2715 (1977). 15. D.F. Mullica, J.D. Oliver, W.O. Milligan and D.F. Hills, J. Inorg. Nucl. Chem., 15, 361 (1979). 16. E. Reguera, J. FernaÂndez-BertraÂn and L. NunÄez, Polyhedron, 13, 1619 (1994).

654 17. P. Gutlich, in U. Gonser (Ed.). MoÈssbauer Spectroscopy, Springer, Berlin, 1975, p. 53 18. E. Reguera and J. FernaÂndez-BertraÂn, J. Radional. Nucl. Chem. Lett., 200, 443 (1995). 19. J. FernaÂndez-BertraÂn, E. Reguera and J. Blanco, Spectrochim. Acta, 46A, 1679 (1990). 20. E. Reguera and J. FernaÂndez-BertraÂn, Eur. J. Solid. State Inorg. Chem., 31, 1021 (1994). 21. E. Reguera, J. FernaÂndez-BertraÂn and J. Duque, Polyhedron, 13, 479 (1994). 22. J. FernaÂndez-BertraÂn and E. Reguera, Solid State Ionics, 93, 139 (1997). 23. J. FernaÂndez-BertraÂn and E. Reguera, Spectrochim. Acta, 49A, 43 (1992). 24. A.R. Champion and H.G. Drickamer, J. Chem. Phys., 47, 2591 (1967).

25. G.J. Lewis and E. Whalley, J. Chem Phys., 47, 2591 (1967). 26. P.G. Sim and E. Whalley, J. Phys. Chem., 91, 1877 (1987). 27. E. Reguera, J. Balmaseda, G. Quintana, A. GoÂmez and J. FernaÂndez-BertraÂn, Polyhedron, 17, 2353 (1998). 28. H. Inoue, T. Nakazawa, T. Mitsuhashi, T. Shirai and E. Fluck, Hyperf. Interact., 46, 725 (1989). 29. G.B. Seifer, Russ. J. Inorg Chem., 5, 33 (1960). 30. J.F. Allen and K. Bonnette, Jr., J. Inorg. Nucl. Chem., 36, 1011 (1974). 31. M.B. Robin, Inorg. Chem., 1, 337 (1962). 32. M.B. Robin and P. Day, Adv. Inorg. Radiochem., 10, 247 (1967).

TMCH 4440